# Does enhanced shape bias improve neural network robustness to common corruptions?

arXiv: https://arxiv.org/abs/2104.09789 · HTML: https://mlsanity.com/paper/2104.09789/

## Key facts

- Cited by 32 papers in the map (data through 14 Oct 2026).
- Builds on 33 of its 39 references that are in the map.
- First posted on arXiv 20 Apr 2021.

## Builds on (most cited first, 30 of 33)

- [Explaining and Harnessing Adversarial Examples](https://mlsanity.com/paper/1412.6572/) (2014)
- [Intriguing properties of neural networks](https://mlsanity.com/paper/1312.6199/) (2013)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features](https://mlsanity.com/paper/1905.04899/) (2019)
- [Benchmarking Neural Network Robustness to Common Corruptions and Perturbations](https://mlsanity.com/paper/1903.12261/) (2019)
- [RandAugment: Practical automated data augmentation with a reduced search space](https://mlsanity.com/paper/1909.13719/) (2019)
- [Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization](https://mlsanity.com/paper/1703.06868/) (2017)
- [The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization](https://mlsanity.com/paper/2006.16241/) (2020)
- [Adversarial examples in the physical world](https://mlsanity.com/paper/1607.02533/) (2016)
- [Group Normalization](https://mlsanity.com/paper/1803.08494/) (2018)
- [ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness](https://mlsanity.com/paper/1811.12231/) (2018)
- [Self-training with Noisy Student improves ImageNet classification](https://mlsanity.com/paper/1911.04252/) (2019)
- [Learning Robust Global Representations by Penalizing Local Predictive Power](https://mlsanity.com/paper/1905.13549/) (2019)
- [Adversarial Examples Are Not Bugs, They Are Features](https://mlsanity.com/paper/1905.02175/) (2019)
- [Holistically-Nested Edge Detection](https://mlsanity.com/paper/1504.06375/) (2015)
- [Using Self-Supervised Learning Can Improve Model Robustness and Uncertainty](https://mlsanity.com/paper/1906.12340/) (2019)
- [Making Convolutional Networks Shift-Invariant Again](https://mlsanity.com/paper/1904.11486/) (2019)
- [A Learned Representation For Artistic Style](https://mlsanity.com/paper/1610.07629/) (2016)
- [Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Coming](https://mlsanity.com/paper/1907.07484/) (2019)
- [A Fourier Perspective on Model Robustness in Computer Vision](https://mlsanity.com/paper/1906.08988/) (2019)
- [Approximating CNNs with Bag-of-local-Features models works surprisingly well on ImageNet](https://mlsanity.com/paper/1904.00760/) (2019)
- [Generalisation in humans and deep neural networks](https://mlsanity.com/paper/1808.08750/) (2018)
- [Domain-Specific Batch Normalization for Unsupervised Domain Adaptation](https://mlsanity.com/paper/1906.03950/) (2019)
- [Physical Adversarial Examples for Object Detectors](https://mlsanity.com/paper/1807.07769/) (2018)
- [Richer Convolutional Features for Edge Detection](https://mlsanity.com/paper/1612.02103/) (2016)
- [Improving Robustness Without Sacrificing Accuracy with Patch Gaussian Augmentation](https://mlsanity.com/paper/1906.02611/) (2019)
- [On Physical Adversarial Patches for Object Detection](https://mlsanity.com/paper/1906.11897/) (2019)
- [Measuring the tendency of CNNs to Learn Surface Statistical Regularities](https://mlsanity.com/paper/1711.11561/) (2017)
- [On Feature Normalization and Data Augmentation](https://mlsanity.com/paper/2002.11102/) (2020)
- [Informative Dropout for Robust Representation Learning: A Shape-bias Perspective](https://mlsanity.com/paper/2008.04254/) (2020)

## Built on by (most cited first, 30 of 32)

- [Intriguing Properties of Vision Transformers](https://mlsanity.com/paper/2105.10497/) (2021)
- [Robust Contrastive Learning Using Negative Samples with Diminished Semantics](https://mlsanity.com/paper/2110.14189/) (2021)
- [On the Effect of Pruning on Adversarial Robustness](https://mlsanity.com/paper/2108.04890/) (2021)
- [Can Biases in ImageNet Models Explain Generalization?](https://mlsanity.com/paper/2404.01509/) (2024)
- [Exploring Corruption Robustness: Inductive Biases in Vision Transformers and MLP-Mixers](https://mlsanity.com/paper/2106.13122/) (2021)
- [ImageNet-trained CNNs are not biased towards texture: Revisiting feature reliance through controlled suppression](https://mlsanity.com/paper/2509.20234/) (2025)
- [Convolutional Neural Networks Rarely Learn Shape for Semantic Segmentation](https://mlsanity.com/paper/2305.06568/) (2023)
- [Trapped in texture bias? A large scale comparison of deep instance segmentation](https://mlsanity.com/paper/2401.09109/) (2024)
- [The Art of Deception: Color Visual Illusions and Diffusion Models](https://mlsanity.com/paper/2412.10122/) (2024)
- [Visual Data Diagnosis and Debiasing with Concept Graphs](https://mlsanity.com/paper/2409.18055/) (2024)
- [Deep Shape-Texture Statistics for Completely Blind Image Quality Evaluation](https://mlsanity.com/paper/2401.08107/) (2024)
- [Robust Asymmetric Heterogeneous Federated Learning with Corrupted Clients](https://mlsanity.com/paper/2503.09206/) (2025)
- [Enhancing 2D Representation Learning with a 3D Prior](https://mlsanity.com/paper/2406.02535/) (2024)
- [Shape Bias and Robustness Evaluation via Cue Decomposition for Image Classification and Segmentation](https://mlsanity.com/paper/2503.12453/) (2025)
- [On the Effects of Artificial Data Modification](https://mlsanity.com/paper/2110.13968/) (2021)
- [Decomposing and Revising What Language Models Generate](https://mlsanity.com/paper/2509.00765/) (2025)
- [Robust Object Detection with Multi-input Multi-output Faster R-CNN](https://mlsanity.com/paper/2111.13065/) (2021)
- [Connecting metrics for shape-texture knowledge in computer vision](https://mlsanity.com/paper/2301.10608/) (2023)
- [Fourier Sensitivity and Regularization of Computer Vision Models](https://mlsanity.com/paper/2301.13514/) (2023)
- [Domain Generalization Emerges from Dreaming](https://mlsanity.com/paper/2302.00980/) (2023)
- [Clarifying Myths About the Relationship Between Shape Bias, Accuracy, and Robustness](https://mlsanity.com/paper/2406.05006/) (2024)
- [Transferring Styles for Reduced Texture Bias and Improved Robustness in Semantic Segmentation Networks](https://mlsanity.com/paper/2507.10239/) (2025)
- [On the Reliability of Cue Conflict and Beyond](https://mlsanity.com/paper/2603.10834/) (2026)
- [Intrinsic Bias Identification on Medical Image Datasets](https://mlsanity.com/paper/2203.12872/) (2022)
- [On the ability of CNNs to extract color invariant intensity based features for image classification](https://mlsanity.com/paper/2307.06500/) (2023)
- [On the Influence of Shape, Texture and Color for Learning Semantic Segmentation](https://mlsanity.com/paper/2410.14878/) (2024)
- [Contrast All the Time: Learning Time Series Representation from Temporal Consistency](https://mlsanity.com/paper/2410.15416/) (2024)
- [A Critical Review of Predominant Bias in Neural Networks](https://mlsanity.com/paper/2502.11031/) (2025)
- [Response Wide Shut? Surprising Observations in Basic Vision Language Model Capabilities](https://mlsanity.com/paper/2507.10442/) (2025)
- [ShapeY: A Principled Framework for Measuring Shape Recognition Capacity via Nearest-Neighbor Matching](https://mlsanity.com/paper/2604.25065/) (2026)

## Abstract (arXiv)

Convolutional neural networks (CNNs) learn to extract representations of complex features, such as object shapes and textures to solve image recognition tasks. Recent work indicates that CNNs trained on ImageNet are biased towards features that encode textures and that these alone are sufficient to generalize to unseen test data from the same distribution as the training data but often fail to generalize to out-of-distribution data. It has been shown that augmenting the training data with different image styles decreases this texture bias in favor of increased shape bias while at the same time improving robustness to common corruptions, such as noise and blur. Commonly, this is interpreted as shape bias increasing corruption robustness. However, this relationship is only hypothesized. We perform a systematic study of different ways of composing inputs based on natural images, explicit edge information, and stylization. While stylization is essential for achieving high corruption robustness, we do not find a clear correlation between shape bias and robustness. We conclude that the data augmentation caused by style-variation accounts for the improved corruption robustness and increased shape bias is only a byproduct.